Position sizing is how much of your capital you put in any single bet. It's not which assets to hold — it's how much. A great idea with 100% of your capital is a single point of failure; a bad idea with 2% is survivable. The same edge, sized too large, can blow up the account; sized too small, it barely moves the needle. The discipline is to match size to edge and to risk: higher conviction and better odds justify larger positions, but only up to a limit that keeps you in the game if you're wrong.
The math is straightforward. If you risk 10% of your capital on one position and it goes to zero, you're down 10%. If you risk 50%, one full loss is catastrophic. Multiple uncorrelated positions reduce the chance of total wipeout — but only if no single position is so large that one loss dominates. Position sizing is the main lever for controlling drawdowns and avoiding ruin. Kelly Criterion and its variants (half-Kelly, fractional Kelly) give a theoretical optimum for how much to bet given edge and odds; in practice, most investors use a mix of rules: max single position, max sector, and a link between conviction and size.
The mental model extends beyond portfolios. In business, position sizing is how much you bet on one product, one geography, or one hire. In time allocation, it's how many hours you put on one project. The principle is the same: size the bet to the edge and to the downside you can tolerate. Over-concentration creates existential risk; over-diversification dilutes edge. The art is in between.
Section 2
How to See It
Position sizing shows up whenever you allocate a limited resource — capital, time, people — across multiple opportunities. Look for: rules that cap single exposure, and the link between conviction and size.
Investing
You're seeing Position Sizing when a fund has a rule: no single position above 10% of the portfolio, and no sector above 40%. A 20% drawdown in one name costs the fund 2%; a 20% drawdown in a 50% position costs 10%. The rule doesn't say which names to pick — it says how much to risk per bet so that one bad call doesn't end the game.
Business
You're seeing Position Sizing when a company limits revenue from any single customer to 15% of total. Losing that customer hurts but doesn't kill the company. The "position" is customer concentration; the "size" is the revenue share. The same logic applies to product lines, geographies, and key people.
Strategy
You're seeing Position Sizing when a founder allocates 60% of engineering to one product bet and 40% to maintenance and experiments. The 60% is a large "position" in one outcome. If the bet fails, the cost is high; if it wins, the payoff is high. Position sizing is the explicit choice of that split — and the decision to cap or increase it as evidence arrives.
Personal
You're seeing Position Sizing when you decide how much of your savings to put in a single asset or how much of your time to give to one side project. Putting 80% of savings in one stock is a huge position; putting 5% in each of 20 ideas is small positions. The tradeoff is concentration (high impact if right) vs survival (low impact if wrong).
Section 3
How to Use It
Decision filter
"Before committing capital, time, or people to a single bet, ask: how much am I putting in? What happens if I'm wrong? Is that size consistent with my edge and my tolerance for drawdown? Cap single positions so that one full loss doesn't ruin you; size up only where edge and conviction justify it."
As a founder
Size your bets. Don't put all engineering into one product unless you're willing to bet the company on it. Don't let one customer dominate revenue. Cap concentration in key people, geographies, and channels so that one loss doesn't force a fire sale. The mistake is "our best idea deserves 80% of resources" — that's not position sizing, that's all-in. The other mistake is spreading so thin that no bet can win meaningfully. Have a max single "position" (customer, product, market) and a logic for how size scales with conviction.
As an investor
Set position limits before you invest. A common rule: no single name above 5–10% of the portfolio; no sector above 25–40%. Adjust for liquidity — you can't size a position larger than you can exit without moving the market. Scale size with conviction: your highest-conviction ideas get the largest positions within the cap. Never size so large that one blow-up wipes out a year of gains or triggers margin or redemption pressure.
As a decision-maker
When evaluating a strategy or a company, look at concentration. A portfolio or business that has 50% in one position is one decision away from a 50% drawdown. Ask: what's the max position? Is it consistent with the stated risk tolerance? Position sizing is often the difference between a strategy that compounds and one that blows up on one bad bet.
Common misapplication: Sizing every position the same regardless of edge. Equal weighting is simple but ignores that some ideas are better than others. Position sizing should reflect conviction and expected payoff — larger where edge is higher and risk is controlled, smaller where uncertainty is high.
Second misapplication: Going "all in" on the best idea. Even a 60% edge can lose. A single position of 100% means one loss is total loss. Position sizing exists so you can be wrong and still play again.
Section 4
The Mechanism
Section 5
Founders & Leaders in Action
Ed ThorpMathematician, author; early quant investor
Thorp used Kelly Criterion for position sizing in blackjack and in his hedge fund. His approach: estimate edge and odds, then size proportionally — but often at half-Kelly or less to reduce drawdowns. He emphasised that position sizing is more important than picking winners — you can have a positive edge and still go broke if you oversize. The discipline is to never bet so much that one bad run wipes you out.
Buffett runs a concentrated portfolio — he's said that diversification is protection against ignorance and that he'd rather put a lot in a few great ideas. But "a lot" is still bounded: Berkshire's largest equity positions are typically in the 5–15% range of the portfolio, and he avoids leverage. So his position sizing is: high concentration relative to the average investor, but with caps that prevent any single position from ending the game. The lesson: concentration is a choice; position sizing is the discipline that makes concentration survivable.
Section 6
Visual Explanation
Position Sizing — How much you put in each bet. Too large: one loss can ruin you. Too small: edge doesn't compound. Size to edge and to the drawdown you can tolerate.
Section 7
Connected Models
Position sizing sits between concentration and diversification, and depends on edge and risk. The models below either prescribe size (Kelly), define the downside (margin of safety, gambler's ruin), or frame the tradeoff (risk-reward, diversification).
Reinforces
Kelly Criterion
Kelly Criterion gives the theoretically optimal bet size as a function of edge and odds. Position sizing is the practice of applying that logic — or a fraction of it — to real bets. Kelly says how much to bet; position sizing adds caps, liquidity constraints, and correlation so you don't bet more than you can survive.
Reinforces
Margin of Safety
Margin of safety is the buffer between your price and the value you need for the bet to work. Position sizing is how much you put in that bet. The two work together: a larger margin of safety can justify a larger position (you're less likely to be wrong), but you still cap size so that if your margin was wrong, you don't get wiped out.
Tension
Concentration
Concentration is putting a lot in few ideas to maximise return when you're right. Position sizing is the discipline that makes concentration safe — you concentrate, but not so much that one loss is fatal. The tension: pure concentration says "all in on the best idea"; position sizing says "cap the size." The resolution is concentrated within limits.
Tension
Diversification
Diversification is spreading risk across many bets. Position sizing can mean many small positions (diversified) or fewer larger ones (concentrated). The tension: diversification reduces single-bet risk but can dilute edge; position sizing can be applied in either direction — more small positions or fewer larger ones, but each sized so no single bet can ruin you.
Section 8
One Key Quote
"Bet sizing is more important than picking winners. You can have an edge and still go broke if you bet too much."
— Ed Thorp, A Man for All Markets
The quote inverts the usual focus. Most people obsess over what to buy; Thorp says the size of the bet often matters more. A modest edge with good position sizing can compound; a great edge with reckless sizing can blow up. The discipline is to size first — then pick.
Section 9
Analyst's Take
Faster Than Normal — Editorial View
Position sizing is the main control for survival. You can't always be right. You can control how much you lose when you're wrong. Cap single positions so that one full loss doesn't end the game. That cap depends on your edge, your risk tolerance, and liquidity — but the principle is non-negotiable.
Size should scale with conviction — within limits. Your best ideas get the largest positions, but even the best idea shouldn't be 50% of the portfolio unless you're willing to lose half your capital on one call. A simple rule: max single position 5–10%, max sector 25–40%. Adjust for your strategy; don't drop the principle.
Concentration and position sizing are not opposites. You can be concentrated (few names) and still size each position so that no one name can kill you. Buffett is concentrated; he's not all-in on one stock. The mistake is confusing "high conviction" with "no cap."
Apply it to customers, products, and people. Revenue concentration in one customer is a position. So is betting 80% of engineering on one product. Cap concentration so that one loss — one churned customer, one failed launch — doesn't force a crisis. The same logic as in a portfolio.
Liquidity constrains max size. You can't have a 20% position in a stock that trades $1M a day when your fund is $100M — you couldn't exit without moving the price. Position sizing in practice has to account for how much you can actually buy and sell. The theoretical Kelly size may be larger than the liquidity-adjusted size; use the binding constraint.
Section 10
Test Yourself
Is this mental model at work here?
Scenario 1
A fund has 30% of AUM in one stock. The stock drops 50%. The fund is down 15% from that single position.
Scenario 2
A company gets 60% of revenue from one customer. The customer signals they may switch vendors next year.
Scenario 3
A portfolio has 20 positions of 5% each. One position goes to zero. The portfolio is down 5%.
Scenario 4
A founder puts 80% of engineering into one new product and 20% into maintenance. The new product fails. The company has little to show for the year.
Section 11
Summary & Further Reading
Summary: Position sizing is how much capital, time, or resource you put in any single bet. It determines how much you lose when you're wrong and how much you gain when you're right. Size too large and one bad outcome can ruin you; size too small and edge doesn't compound. The discipline is to cap single positions so that one full loss is survivable, and to scale size with conviction within those caps. Applies to investing (portfolio weights), business (customer and product concentration), and any allocation of limited resources across uncertain outcomes.
Mathematical basis for optimal bet sizing; half-Kelly and fractional Kelly as practical variants.
Leads-to
[Gambler's Ruin](/mental-models/gamblers-ruin)
Gambler's ruin is the probability of going broke with repeated bets. Position size directly determines that probability: bet too much per round and ruin approaches 1; bet small enough and you can survive variance. Position sizing is the main lever to stay below the ruin threshold.
Leads-to
Risk-Reward [Ratio](/mental-models/ratio)
Risk-reward ratio frames the upside vs downside of a bet. Position sizing multiplies that: a 2:1 risk-reward bet with 20% of capital has a different impact than the same bet with 5% of capital. Size determines how much you actually risk; risk-reward determines the shape of the bet. Use both.
